The goal of this project is to advance our understanding of the neuroanatomical localization of syntactic processing in sentence comprehension. We will investigate this question by correlating deficits in syntactic comprehension in patients with unilateral strokes with the location and size of brain lesion as determined by structural MR and metabolic activity (PET). The study will use a series of off-line and on- line measures to study disorders of syntactically-based sentence comprehension. A sophisticated, topographic landmark based parcellation system that has been designed and implemented at M.G.H. will be used to analyze structural MR images. This parcellation system will be extended to co-registered PET datasets. The image analyses will provide a description of the entire hemispheric extent of a lesion in terms of regions of hypometabolism as well as necrosis. The type and severity of syntactic disorder will be correlated with the location and extend of lesions in cortical regions, subcortical gray matter nuclei, and what matter tracts. The research will make contribution in several areas. The results of the psycholinguistic investigations of syntactic comprehension disorders will produce new data regarding the nature of these impairments. The neuroimaging data will explore the relationship of areas of hypometabolism shown by PET to the extent of necrosis seen on MR, and provide data on the role of deafferentation in the determination of cortical hypometabolism. The combination of powerful morphometric analyses with detailed assessment of very specific impairments of language function will allow us to investigate the localization of specific aspects of syntactic comprehension in discrete brain regions at a level of detail not previously possible. The data will provide important new knowledge about the location of lesions that produce these deficits, and will be relevant to models of the functional neuroanatomy of this aspect of language processing.